[no subject]

I think a better solution if were known to work would be something
that only slowed down from a serial interrupt to an rti, but
that's probably not good enough, and doesn't handle serial sends
I think.

The whole situation is hard to solve. Some code has to be speed
dependent and on the IIGS it is hard to impossible to write speed
independent code for some i/o (assuming a certain lack of status
info available from the i/o device, which I assume is the case for
some devices).
--
-- Jay    jay.krell@cornell.edu    http://crux4.cit.cornell.edu/~jmk3 --